Linkage 7/15: Chest Scan Caution & Under the Influence of Flags

alpha_1-antitrypsine_deficiency_lung_ct_scanCancer used to be a black box, a disease that physicians could only monitor through surgical biopsies and indirect measures. But for the last thirty years, the use of computed tomography imaging, better known as CT scans, has allowed oncologists and cancer researchers to keep close watch on the growth or shrinkage of a tumor for many different types of cancer. A patient with a lung tumor, for example, can be scanned every few months in order to see whether their therapy is working – and if it’s not, doctors may choose to switch treatments. Clinical trials of new therapies for cancer also make use of CT scans, using the increase or decreased size of the tumor as a primary data point.

But for all the benefits of scans over surgeries to monitor tumor size, flaws remain for CT scans. A new study published this week in the Journal of Clinical Oncology shines a harsh light on one of the primary problems – the technology’s variability. Patients usually are given CT scans months apart, and trained radiologists measure the tumors to see whether they are growing or receding. But how much of those changes can be attributed to random error from the imperfect resolution of the scan or the breathing of the patient?

To test this baseline error, researchers from Memorial Sloan-Kettering Cancer Center got a little tricky. Instead of taking two scans from a patient months apart, they took two scans in quick succession, within 15 minutes. The scans were then handed off to experienced radiologists, who were told to measure the change in tumor size without knowing how much time had elapsed between the images. The results were sobering – despite the tumor being biologically identical between the two near-simultaneous scans, the radiologists found changes in size of 1mm or more in more than half of the samples and a 10 percent error range in either direction overall. Although the criteria for tumor progression is an increase in size of 20 percent or more, that 10 percent error could considerably distort the data when clinical and research decisions are made using normally-spaced scans.

The result doesn’t render CT scans obsolete, but offers new caution about the method’s shortcomings.

“It’s the sense of, ‘Really? Is this first happening now?'” Michael Maitland, assistant professor of medicine at the Medical Center, commented to Reuters Health about the study findings. “This is telling us scientifically how much noise is naturally there without any treatment or the cancer getting worse. It’s an important thing to do whenever you are going to use any kind of marker for a disease.”

In an accompanying editorial in the Journal of Clinical Oncology, Maitland went further, writing with his co-authors that it was time for oncologists to rely less upon CT scans alone and move toward integrating those images with other measures to create more precise monitoring technologies. As cancer edges toward more personalized treatment strategies, developing better diagnostic tools will become even more important, they argued.

“It is time to cast away familiar conventions and turn to better methods of evaluating malignant disease therapeutics,” they wrote. “It is time to replace these systems with more innovative, quantitative approaches that have the potential to define relationships between solid tumors, disease progression, and therapeutic outcomes in patients.”


It might have come out a few days late for the 4th of July, but Travis Carter’s study of the effects of seeing the American flag on political beliefs is still timely. If the Booth Business School researcher is right, we’ll all be slightly more Republican for at least the next 8 months. Ed Yong at Not Exactly Rocket Science did a great writeup that was featured on the Colbert Report this week (and also wrote up our own Neil Shubin’s study on the origin of limb genetic programs this week as well).

Childhood obesity is a growing problem in the United States, but is it a justifiable reason to remove kids from their parents’ custody? An editorial in JAMA takes this controversial position by suggesting that it may be more ethical to place obese children in foster care than to give them surgery. “State intervention may serve the best interests of many children with life-threatening obesity, comprising the only realistic way to control harmful behaviors,” write authors Lindsey Murtagh and David Ludwig. University of Chicago physician and ethicist Lainie Ross commented on the idea to the Associated Press.

To visualize complex science, former Medical Center employee Matthew McCrory draws upon experience in both hospitals and Hollywood. A nice New York Times article on science art profiles McCrory and his current work at Northwestern University.

About Rob Mitchum (525 Articles)
Rob Mitchum is communications manager at the Computation Institute, a joint initiative between The University of Chicago and Argonne National Laboratory.
%d bloggers like this: